Throughout history, placebo effects have been variously considered as tricks played upon the gullible by medical practitioners and powerful but mysterious healing forces. With the advent of direct measurements of human brain function, modern science has shown that placebo effects are neither of these. Rather, they reflect the principled impact of psychological and brain processes on diseases of the brain and body. Placebo effects represent an opportunity-because they provide a window into internal brain processes that influence health- and a challenge, because dozens of clinical trials have now failed to show the superiority of commonly prescribed drugs to placebo, at great cost to health care providers and consumers. This may not be due to the ineffectiveness of the medications, but rather seems to be related to high placebo response rates. Studies of the brain mechanisms underlying placebo effects can provide valuable insight into how the brain actively anticipates, responds to, and learns about active treatments, and the effects of these processes on pain and other aspects of health. In the previous period of this grant, functional magnetic resonance imaging (fMRI) of placebo and opiate drug effects was instrumental in establishing that placebo treatments impact the neurobiology of pain, including effects on the cortical and subcortical brain regions most closely associated with pain experience. The research also provided some of the first evidence on the interactions in the brain that give rise to placebo effects, and demonstrated the involvement of prefrontal cortical-subcortical-brainstem systems and the endogenous opioid system. The view that emerged is that placebo treatments engage brain systems related to both conceptual thought (explicit beliefs) and learning processes that change the value assigned to treatment cues and pain itself. In this Renewal application, we build on this foundation to address some of the many unanswered questions that need to be addressed before placebo effects are understood, harnessed in therapy, and separated cleanly from drug effects in clinical trials and practice. These include fundamental questions about a) the nature of the pain-related processes regulated by placebos (are they specific to pain, or general affective processes?) and b) the critical conditions required to create placebo analgesia. Our working hypothesis is that conceptual and learning processes interact, and that both are critical. Learning systems fundamentally shape the motivational value of pain and modulate nociception in subcortical-brainstem systems, and conceptual processes are needed to attribute perceived benefits to the placebo treatment and engage learning. To better understand these systems, we bring to bear new computational machine-learning tools that can identify and test placebo effects on patterns of brain activity specific to pain. We also test the involvement of several neurochemical systems-opioids, dopamine, and oxytocin-and their relationships with specific brain and psychological aspects of the placebo response-generation process.